Chaos Engineering For Hybrid Targets With LitmusChaos

Litmus Chaos is a tool for chaos engineering that focuses on making it more robust and lean, and on enabling developers to use it earlier in the development process. The tool has monthly releases and has added new fault types, support for newer versions of Kubernetes and OpenShift, and support for randomization of fault inputs.

• Litmus Chaos is a tool for chaos engineering that focuses on making it more robust and lean
• The tool has monthly releases and has added new fault types
• Support for newer versions of Kubernetes and OpenShift has been added
• Support for randomization of fault inputs has been added
• The tool is being designed to enable developers to use it earlier in the development process

The speaker demonstrated how to use the Grafana dashboard to monitor the behavior of the application during a fault injection. They also showed how to use the Litmus API to automate the process and how to compare different runs of the experiment to see if resilience has improved. They also mentioned that the tool has had contributions from over 200 different contributors from various companies.

Downtime amongst platform resources is very common today and is one of the reasons why SRE teams are heavily invested in Chaos Engineering. Another upcoming phenomenon is that of multi-cloud deployments wherein organizations attempt to achieve redundancy through cross-cloud instances of their services. These requirements together warrant a vendor-agnostic approach to chaos experimentation, preferably from a common control plane (read: Kubernetes) with the flexibility to add custom constraints around the fault injection process as well as the expectations (hypothesis) around them. There is also a need to marry this process with the standard CI/CD pipelines to derive maximum benefit via continuous chaos testing. In this session, the maintainers of the LitmusChaos project explain how this can be achieved.